Source: https://www.reddit.com/r/launchschool/comments/1ujidfs/2025_capstone_salary_data/

Bootcamps died in 2025 and I got feedback that continuing to monitor them so closely in 2026 is like "beating a dead horse".

26 minutes ago Launch School released their numbers and I'm reporting on it here.

I believe in transparency and integrity and we've all seen what happens to institutions that do not... it eventually catches up with them.

Well here we are with Launch School's 2025 outcomes data.

As usually the original data is so detailed I have FAR LESS commentary then I do on a bootcamp like Codesmith that cares more about manipulating the story about them than delivering good results. If your outcomes are transparent, good or bad, then it's easy. If your outcomes are sketchy and manipulative, then it's comment wars and anonymoous paid attack accounts.

ANALYSIS:

• The denominator matters. Launch School lays out three numbers, which goes to show how other bootcamps that quote the best number on their website are fudging definitions to make themselves look good. Across all enrollees, there is basically a ~58% six-month placement story from a program that used to have a 100% placement rate. But, like CIRR, it's fair to remove people who didn't intend on job hunting in the first place, which is where I get the 65% rate. But you can also see how adding in "internships" turns that into 70%, which is something that CIRR might also allow, but I think is a tad too far in this context.
• 65% is quite strong. Launch School's closest competitors have been Rithm (which shut down in 2024) and Codesmith (which pivoted to become "Codesmith Enterprise" and I haven't seen a single placement in months).
• Remote was the rug pull that took the bootcamp industry with it. A few years ago, the pitch was effectively: learn from anywhere, get a remote software job, change your life. That world is mostly gone. Remote placements going from ~78% to ~27% is not a small shift. It changes who can realistically succeed, what kind of job search people need to run, and whether “online bootcamp” still maps cleanly to the actual market.
• Location is part of the curriculum now, whether bootcamps admit it or not. The uncomfortable truth is that being near a bit tech city seems to matter a lot again. Not because geography magically makes you smarter, but because startups, in-person roles, founder networks, FDE jobs, and apprenticeship-style opportunities are clustering there. A bootcamp that treats location as irrelevant is probably giving students outdated advice.
• AI didn’t make fundamentals obsolete; it made fake fundamentals easier to expose, which is why Launch School’s approach matters. They haven’t slapped “AI-first” on a bootcamp or pivoted 100% into hype; they’ve woven AI through a real engineering foundation so students learn to read code, reason about systems, debug hard problems, and use AI as leverage, not a substitute. This is smart and takes a lot of judgement and taste to do right.

Details below. Interesting to see Gauntlet AI take the 10-week training program.

1. Fedstack (Smoothstack): $1,623,840.00

• Task Order ID: 2032L226F00064
• Fedstack secured the highest-value award to scale highly governed training pipelines that transition existing personnel into specialized AI engineering roles.
• https://orangeslices.ai/fedstack-secures-department-of-the-treasury-artificial-intelligence-engineer-workforce-development-task/

2. Gauntlet AI (BloomTech): $1,420,654.80

• Task Order ID: 2032L226F00062
• Gauntlet AI received funding to deploy intensive 10-week fellowships and hybrid upskilling tracks aimed at teaching federal tech teams how to securely build and deploy production-grade Large Language Model (LLM) applications.
• https://orangeslices.ai/gauntlet-wins-department-of-the-treasury-artificial-intelligence-engineer-workforce-development-task/

3. Fearless Solutions: $756,000.00

• Task Order ID: 2032L226F00061
• Fearless Solutions was awarded this task order to train federal developers on seamlessly integrating AI utilities and automated testing into standard agile software workflows and continuous deployment pipelines.
• https://orangeslices.ai/fearless-wins-department-of-the-treasury-artificial-intelligence-engineer-workforce-development-task/

4. SoKat Consulting: $552,360.00

• Task Order ID: 2032L226F00063
• Scope: SoKat Consulting secured funding to deliver specialized training focused on advanced data stewardship, secure custom LLM assistants, and regulatory-compliant agentic workflows.
• https://orangeslices.ai/sokat-beats-out-3-to-win-department-of-the-treasury-artificial-intelligence-engineer-workforce-development-task/

Developing.

End of an era.

The new website has four pages: homepage, process page, case study page, and a contact form.

The company is calling itself "Codesmith Enterprise" now. There are no legally required terms of service or privacy policy to identify if the company structure has changed.

The new tagline is:

>Building AI Capability, Piece by Piece. We move enterprise and government teams from pilot discovery to engineering capacity that produces measurable results as AI advances.

The website highlights their IRS training work, and solicits contact for enterprise, government, talent sourcing, and up-skilling.

----

We haven't heard official word from Codesmith on what's happening and their most recent releases promised a bright future and that they were going to be around for 10 more years.

These are on Course Report's "top bootcamps of 2026" list that do not have a single review IN 2026 and I'm struggling to find people who actually went there in 2026.

Codesmith (excluding Future Code)

Flatiron

Coding Temple

Codeworks

Icon Hack

Codesmith's "free learning platform" that they claim can get you an "entry level job" by completing is one of the worst pieces of software I've seen in ed-tech.

For example, this URL gives the ENTIRE CONTENT IN ITS ENTIRETY away in a single click: https://csx.codesmith.io/api/yaml/all

This isn't a secret or a hack, it's literally just fetched when you login to load all the content and is freely available to the internet, no authentication.

Next, anyone can just grab all the user data for any user they want.

For example, this is my user data: https://csx.codesmith.io/users/95526 and I give you permission to open it and look at it.

DO NOT look at other users data because that can be a gray area if you scrape it, even if Codesmith doesn't have any permissions or controls on their own.

Enough is enough here, no pretending and shut it down.

P.S. Will Sentance, you need to take responsibility for this level of garbage software and Codesmith students deserve to know if their personal information has been at risk.

I asked CIRR 30 days ago where the 2024-25 missing reports are. I did not get a response. This tells me CIRR is dead and its flagship Codesmith is dead. Other bootcamps we've lost are: Rithm, Turing, Codeup, Kenzie, Launch Academy, Momentum, Alchemy, Epicodus, Lighthouse Labs, 2U/Trilogy, Lambda School, and more.

Unlike the embarrassing end that CIRR and Codesmith are experiencing - too ashamed to end on a positive note and instead end in layoffs and utter silence, I want things to end on a positive.

If you graduated from a coding bootcamp in the past, and it changed your life, TELL US YOUR STORY. No selling or shilling, just tell us how coding impacted your life and in the right time and right place your bootcamp experience mattered.

An SEO guy, Lars Lofgren, wrote what I think was an unfact-checked hit piece about me.

It had a bunch of trivial and self-contradicting factual mistakes, left out key context, and got shared around by other SEO people anyway.

I wrote an 8000 word response walking through the errors, debunking the piece, and setting the record straight, but the whole thing left me genuinely confused.

It feels like I'm playing a game I don't understand.

Is this normal in SEO?

In my world, if you publish something damaging about a real person, you fact-check it. You reach out. You separate facts from assumptions. You don’t just write the spiciest version and let the target clean up the mess later.

So I’m asking honestly: is the SEO community generally okay with “viral first, facts later” content? Or is this considered bad form by serious people in the industry?

Trying to understand how to navigate this world.